lecture 1: labour economics and wage-setting...

Lecture 1: Labour Economics and Wage-Setting Theory

Spring 2015

Lars Calmfors Literature: Chapter 1 Cahuc-Zylberberg (pp 4-19, 28-29, 35-55)

The choice between consumption and leisure U = U(C,L) C = consumption of goods L = consumption of leisure L0 = total amount of time h = L0 – L = working time U(C,L) = U defines an indifference curve Figure 1.1

U(C,L) = U defines a function C(L), which satisfies U[C(L),L] = U

1

2

Differentiation w.r.t L gives:

0'C L

U C U+ =

( , )

'( ) ( , )

L

C

U C LC L

U C L= −

,

( , )( )

( , )' L

C L

C

U C LC L MRS

U C L= =

Indifference curves are negatively sloped. Indifference curves are convex (absolute value of slope falling with L) if C″(L) > 0. C″(L) is obtained by differentiating C′(L) = -UL(C,L)/UC (C,L) w.r.t L and substituting -UL/UC for C′ after differentiation.

3

We get:

C″(L) = 2

( )

2 C L

CL LL CC

L C

C

L

U

U UU U

U UU U

− −⎡ ⎤⎢ ⎥⎢ ⎥⎣ ⎦

C″(L) > 0 if 2 0C L

CL LL CC

L C

U UU U U

U U− − >

This is certainly the case if

CLU = 0 since

LLU < 0 and

CCU < 0.

4

The choice problem of the individual w = real hourly wage wh = real wage income R = other income The individual’s budget constraint: C ≤ wh + R Alternative formulation of budget constraint: C ≤ w(L0 – L) + R C + wL ≤ wL0 + R ≡ R0 Interpretation:

• The individual disposes of a potential income R0 obtained by devoting all of his time to working and using other resources R. Leisure or consumer goods can be bought with this income.

• The wage is the price as well as the opportunity cost of leisure. The decision problem of the individual: Max U(C,L) s.t. C + wL ≤ R0 {C,L}

5

Interior solution, such that 0 < L < L0 and C > 0. µ > 0 is the Lagrange multiplier. The Lagrangian is:

£(C,L,µ) = U(C,L) + µ(R0 - C - wL) The FOCs are: Uc (C,L) - µ = 0 UL (C,L) - µw = 0 The complementary slackness condition: µ(R0 - C - wL) = 0 with µ ≥ 0 Since µ = Uc (C,L) > 0 with an interior solution, it follows that the budget constraint is then binding, i.e. C + wL = R0 The optimal solution is then:

0

( *, *) *

( *, *)

* *

L

C

U C Lw

U C L

C wL R

=

+ =

6

Figure 1.2

Equation of budget line: C + wL = R + wL0 = R0 C = R + w(L0 – L) L = L0 ⇒ C = R L = 0 ⇒ C = R + wL0 = R0

• Change in w rotates budget line around A • Change in R gives rise to a parallel shift of the budget line

7

The reservation wage

• E must lie to the left of A for there to be a positive laboursupply (L<L0)

1. Tangency point at A: L = L0 and h = L0 - L = 0 is interior solution

2. Indifference curve is more sloped than budget line at A: L = L0and h = L0 - L = 0 is a corner solution

3. Indifference curve is less sloped than budget line at A: L < L0and h = L0 - L > 0 is an interior solution

E

A

L0 L

C

8

MRS at point A is called the reservation wage, wA

wA =0

0

( , )

( , )L

C

U R L

U R L

• An individual participates in the labour force only if w > wA.

• The reservation wage depends on non-wage income.

• If leisure is a normal good (i.e. MRS increases with income), then ahigher non-wage income creates a disincentive for work.

10

We then obtain:

1

2

CL C CC L C

C

L L

C C

CL LL CC

L L

U U U U UL U

U UL

U UwU U U

U U

∗

−− −

∂Λ = =

∂− −

⎛ ⎞ ⎛ ⎞⎟ ⎟⎜ ⎜⎟ ⎟⎜ ⎜⎟ ⎟⎜ ⎜⎝ ⎠ ⎝ ⎠⎡ ⎛ ⎞ ⎤⎟⎜⎢ ⎥⎟⎜ ⎟⎜⎝ ⎠⎢ ⎥⎣ ⎦

*

0

2

2

C L

CL LL CC

L C

CL C CC L

LL

R U UU U U

U U

U U U U

U∂Λ = =

∂− −

−

⎡ ⎤⎛ ⎞⎛ ⎞ ⎟⎟⎢ ⎜ ⎥⎜ ⎟⎟ ⎜⎜ ⎟ ⎟⎜ ⎜⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦

• From quasi-concavity (convex indifference curves) we have that the denominators of ∧1 and ∧2 are positive.

• Hence signs of ∧1 and ∧2 are determined by the numerators.

• ∧2 > 0 if UCL UC - UCC UL > 0. This is the condition for leisure to be a normal good, i.e. for leisure to increase if income increases.

• ∧1 < 0, i.e. leisure falls and labour supply increases if the wage increases, unambiguously only if leisure is a normal good.

• There is both an (indirect) income effect and a substitution effect. Both are negative if leisure is a normal good.

11

The effect of an increase in non-wage income R:

Figure 1.2

12

The effect of a wage increase

0 0 0

*0

1 2 1 2 0

( ) ( )

* ( , )

L w R R wL R

RdLL

dw w

− +

= = +

∂= + = +

∂∧ ∧

∧

∧ ∧

13

Figure 1.3

• w increases from w to w1

Keep R0 unchanged. New budget line A1R0. As if decline from R to Rc = R – (w1 – w)L0.

Rc = compensated income. A1R0 is the compensated budget constraint.

1. E →E′ is substitution effect reducing leisure. (Outlays ofthe consumer are minimised under the constraint ofreaching a given level of utility.)

2. E′ →E″ is (indirect) income effect reducing leisure fartherif leisure is normal good.

14

3. E″ →E1 is (direct) income effect increasing leisure ifleisure is a normal good. It represents the increase inpotential income from the wage increase.

Conclusion: Net effect of a wage increase on leisure/hours worked is ambiguous.

Simpler analysis:

1. E →E1 is substitution effect

2. E′ →E1 is global income effect (the indirect and direct incomeeffects are aggregated)

Compensated and uncompensated elasticity of labour supply

h = L0 – L* = ∧(w, R0) is the Marshallian (uncompensated) labour supply.

The Hicksian (compensated) labour supply is obtained as the solution to the problem:

Min C + wL s.t. U(C,L) ≥ Ū L,C

One then obtains L = L (w, Ū)

17

Figure 1.4

L0 – L = h

18

Complications

• Higher overtime pay• Progressive taxes• Fixed cost to enter the labour market• Only jobs with fixed number of hours

L0 - Lf = h0 is the fixed number of hours demanded.

19

Figure 1.5

• E is the unconstrained optimum.

• If E is to the left of Ef, the individual would have liked to supplymore hours.

• If E is to the right of Ef, the individual takes the job only if Ef is tothe right of EA (i.e. offering higher utility). The individual is forcedto work more than he would want.

• If Ef is to the left of EA, the individual chooses not to work. Voluntary non-participation.

The condition for taking a job is:

[ ]

[ ]

0 0

0 0

( ), ( , )

( ), ( , ) defines the reservation wage .

f f

A f f A

U R w L L L U R L

U R w L L L U R L w

+ − ≥

+ − =

Utility of working with reservation wage = Utility of not working

21

Aggregate labour supply and labour force participation

• Aggregate labour supply is obtained by adding up the totalnumber of hours supplied by each individual.

• The existence of indivisibilities in working hours offered to agentsimplies that the elasticity of aggregate supply differs from that ofthe individual supply.

• Reservation wages differ among individuals

- differences in preferences - differences in non-wage income

• The diversity of reservation wages [0, ]A

w ∈ +∞ is represented by the cumulative distribution function ( )wφ .

• ( )wφ represents the participation rate, i.e. the proportion of thepopulation with a reservation wage below w.

• If the population size is N, ( )wφ N is the labour force.

• Given N, the wage elasticity of the labour force is equal to that of the participation rate.

• The elasticity is positive, since a higher wage draws workers intothe labour market.

• Key empirical result: the wage elasticity of the participation rate ismuch larger than the wage elasticity of individual labour supply.

23

Labour supply with household production U = U(C,L)

C = CD + CM

CM = quantity of consumption goods bought in the market

CD = home production of consumption goods L0 = total endowment of time

hM = working hours in the market

hD = working hours in the household production

L = leisure L0 = hM + hD + L Home production function: CD = f(hD) f′ > 0, f″ < 0 whM = wage earnings R = non-wage income

Choose CM, CD, hD, hM and L such that utility is maximised s. t. CM ≤ whM + R

CM ≤ whM + R hM = L0 - hD – L ⇒ CM ≤ w(L0 - hD - L) + R CM + wL ≤ wL0 - whD + R

24

wL0 + R = R0 ⇒ CM + wL ≤ R0 - whD

0 + + + -

M D D D

C

C C wL R C wh≤ C + wL ≤ R0 + [f(hD) - whD]

The consumer’s programme Max U(C,L) s.t. C + wL ≤ [f(hD) - whD] + R0 C,L, hD

According to the budget constraint, the total income of the consumer is equal to the sum of potential income R0 and “profit” from household production, f(hD) - whD. Two-step solution Step 1: Choose hD so as to maximise profit from household production and thus also total income:

*'( ) D

f h w= Step 2: Given hD, equivalent problem to that of the basic consumption/leisure model

• Replace

* *

0 0 0 0

* *

0

= + by ( )

= ( )

D D

D D

R wL R R R f h wh

wL R f h wh

= + − =

+ + −

25

The optimal solution is then defined by:

* *

* * *

0* *

( , ) '( ) and

( , )L

D

C

U C Lw f h C wL R

U C L= = + = (5)

Interpretation: • Marginal rate of substitution between consumption and leisure is

equal to the wage.

• Use time for household production up to the point when the marginal productivity of household production = the wage.

• The wage elasticity of labour supply is affected by the possibility to make trade-offs between household and market activities.

(5) gives: *

0 ( , )L w R= ∧

Differentiation w.r.t w:

*

0

1 2

dRdL

dw dw= ∧ + ∧ with

*0

0 D

dRL h

dw= −

26

Since * * *

0

M Dh L h L= − − we have:

* * *

M Ddh dh dL

dw dw dw= − −

Since *

*

* 1 '( we have 0

"( )) D

D

D

dhw f

dw f hh= = <

Using that, we obtain

*

*

0*

0 *

1 2

*1 2 2

1

( ) "( )

1 ( )

"( )

M

D

D

D

D

L

dhL h

dw f h

f hh

= ∧ ∧ − =

= ∧ ∧ ∧

− − −

⎡ ⎤⎢ ⎥− + + −⎢ ⎥⎣ ⎦

–

01 2( ) L∧ ∧+ is the impact on labour supply given household production: ambiguous sign.

*

*2

1

"( )

D

Df h

h∧ − is unambiguously positive if leisure is a

normal good ( )2 0 ∧ > . The possibility to make trade-offs between household production and market work increases the wage elasticity of labour supply.

28

Intrafamily decisions

Interdependent decisions within a family The unitary model

• Extension of the basic model

• Utility of the family is U = U(C, L1, L2) C = total consumption of goods of the family Li (i = 1,2) = leisure of individual i Utility from consumption does not depend on distribution of consumption.

Programme of the household: Max U(C, L1, L2) C, L1, L2 s.t. C + w1L1 + w2L2 ≤ R1 + R2 + (w1 + w2)L0

• Distribution of non-wage incomes does not matter, only their sum

R1 + R2 (income pooling).

• Empirically questionable - Fortin and Lacroix find support only for couples with pre-school- age children.

29

The collective model • Household choices must arise out of individual preferences

• But Pareto-efficient decisions Programme:

Max U1(C1, L1) C1, C2, L1, L2

s.t. 2 2 2 2( , )U C L U≥

C1 + C2 + w1L1 + w2L2 ≤ R1 + R2 + (w1 + w2)L0

2U likely to depend on wi and Ri .

Chiappori (1992):

Max Ui(Ci, Li) s.t. Ci + wiLi ≤ wiL0 + Φi Ci, Li

• Φi is a sharing rule such that Φ1 + Φ2 = R1 + R2

Φi depends on wi and Ri

• Efficient allocations are solutions to individual programmes where each individual is endowed with a specific non-wage income which depends on the overall income of the household.

• Also extensions of basic model with specification of the individual’s non-wage income.

30

Models of intrafamily decisions

• Explanation of specialization in either household or market work

• Interdependence of decisions - w ↓ ⇒ reduction in household income ⇒ increased participation (from earlier non-participants)

- but this additional worker effect does not seem empirically important

- not negative but positive relationship between participation and average wage

31

Empirical research ln h = αw ln w + αR ln R + xθ + ε R = measure of non-wage income x = [x1, …..xn] = vector of other determinants θ = = vector if parameters ε = random term Problem: How to define R. Rt = rtAt-1 + Bt

rt = real rate of interest

At-1 = assets

Bt = exogenous income • This formulation assumes myopic behaviour

• More reasonable to assume intertemporal decisions

• More complex model is required

32

Empirical results • Variations in the participation rate (extensive margin) are

more important for labour supply than variations in working time (intensive margin)

• Female labour supply is much more elastic than male labour supply

• Hump-shaped labour supply as predicted by theory

• Leisure is a normal good

• Substitution effect dominates income effect of wage change for working time

• Only substitution effect for participation rate

33

34

35

Natural experiments and difference-in-differences estimators Population of size N NM has been affected by policy change. NC is control group which has not been affected. δit = 1 if policy change applies to an individual δit = 0 if policy change does not apply to an individual yit = αδ it + x it θ + γ i + ξ t + ε it (21) γ i = individual fixed effect

ξ t = fixed time effect

ε it = random term distributed independently among individuals

x it = vector of observable characteristics

Eliminate individual fixed effects by estimating equation in differences:

Δyit = αΔδ it + (Δx it)θ + Δξ t + Δε it

• Two periods • Same treatment for all in t -1 • Different treatment in t • Assume Δx i = 0 • Set β = Δξ t and ui = Δε it Δyi = β + αΔδ i + ui

37

Value and limits of natural experiments

• Methodological simplicity

• Few situations

• Particular event which perhaps cannot be generalised

• Social experiments